Phosphomimetic Mutants of Pigment Epithelium-Derived Factor with Enhanced Anti-Choroidal Melanoma Cell Activity In Vitro and In Vivo

Abstract

Purpose.: Currently, choroidal melanoma is chemoresistant and there is no routine adjuvant chemotherapy for it. We investigated whether pigment epithelium-derived factor (PEDF) and its triple phosphomimetic mutants could more efficiently suppress melanoma tumor growth and metastasis, as well as how the triple phosphomimetic mutants act as antitumor agents.

Methods.: Phosphomimetic mutants of PEDF were constructed by site mutagenesis. Lentiviruses carrying wild type (WT) PEDF, S24E114E227A (EEA)-PEDF, and S24E114E227E (EEE)-PEDF were produced in 293 fast-growing, highly transfectable (FT) cells and used to infect human choroidal melanoma cell line (OCM-1). The growth, migration, invasion and metastasis abilities of OCM-1 cells expressing WT-PEDF, EEA-PEDF or EEE-PEDF were investigated in vitro and in vivo, while the underlying mechanism of PEDF phosphomimetic mutants were investigated via Western blotting.

Results.: OCM-1 cells infected with lentiviruses carrying WT-PEDF, EEA-PEDF, and EEE-PEDF displayed reduced proliferation, migration and invasion abilities, and were more prone to apoptosis. Cell media containing WT-PEDF, EEA-PEDF, or EEE-PEDF protein inhibited the tube forming capacity of human umbilical vein endothelial cells (HUVEC) in vitro. OCM-1 cells expressing WT-PEDF, EEA-PEDF, or EEE-PEDF displayed significantly reduced tumor growth and metastasis in the melanoma xenograft of nude mice models, with the PEDF mutants displaying much stronger effects than the wild type. The antitumor effects of PEDF are associated with the inhibition of VEGF and nuclear factor kappa-B (NF-κB) expression, as well as further inhibition of Akt phosphorylation.

Conclusions.: The phosphomimetic mutants of PEDF showed enhanced anti-melanoma activity by directly affecting tumor cells and indirectly affecting angiogenesis. These findings encourage the development of PEDF mutants as innovative anticancer agents.

Introduction

Choroidal melanoma, a subtype of uveal melanoma, is the most common type of primary malignant intraocular tumor in adults, accounting for 70% of all primary intraocular eye cancers, and the second most common type of primary malignant melanoma. As for the management of choroidal melanoma, there has recently been a shift from enucleation to more conservative treatments, including thermotherapy, plaque radiotherapy, charged-particle radiotherapy, local resection, and systematic chemotherapy. ¹ However, choroidal melanoma is a chemoresistant tumor, and approximately 50% of patients die within 10 years. Additionally, the median survival time is only 5 to 7 months following the development of metastasis. ² So far, no adjuvant chemotherapy is available as a routine treatment. 

Pigment epithelium-derived factor (PEDF) was first purified from human retinal pigment epithelial (HRPE) cell conditioned media as a factor with potent neurotrophic and differentiative activity. ^3,4 It has also been shown to be a potent inhibitor of angiogenesis in the mammalian eye by inhibiting endothelial cell growth and migration. ^5,6 It is widely expressed throughout the human body and can be detected in systemic circulation as well. ^7,8 PEDF expression is inversely correlated with cancer progression, ⁹ intratumor microvessel density (MVD), ^10,11 metastatic potential, ^11,12 and unfavorable prognoses ^10,11 ; PEDF-overexpressing cancer cells exhibit a reduced growth rate in vivo. ^13,14 Therefore, in preclinical studies, PEDF has been widely tested as an anticancer agent. ^15–18  

PEDF is a glycoprotein that belongs to the superfamily of serine protease inhibitors. It can be phosphorylated by casein kinase 2 (CK2), on serine 24 and serine 114, and by protein kinase A (PKA), on serine 227. ^19–21 It has been reported that differential phosphorylation can modify the function of PEDF, and, therefore, contributes to the complexity of PEDF action. ^20,21 Recently, Konson et al. ²² produced two phosphomimetic mutant PEDF proteins that possess stronger antitumor activities than the wild-type PEDF. 

PEDF suppresses tumor growth through the inhibition of tumor angiogenesis, which involves inhibition of activity and/or expression of VEGF. ^23–26 More recent studies show that PEDF may also possess direct antitumor effects that either inhibit proliferation or promote differentiation of tumor cells. ^27–29 However, the underlying molecular mechanism by which PEDF causes tumor suppression is not yet completely understood. 

In the present study, we investigated whether constitutive expression of PEDF, or its triple phosphomimetic mutant, in choroidal melanoma tumor cells could more efficiently suppress melanoma tumor growth and metastasis, as well as how the triple phosphomimetic mutants act as antitumor agents. The results from our study revealed novel beneficial aspects of PEDF's triple phosphomimetic mutants' effects on melanoma cells. For example, the mutants directly suppress the proliferation, migration, and invasion abilities of melanoma cells, and induce apoptotic cell death as well. PEDF also inhibited tumor angiogenesis via suppression of VEGF expression in tumor cells and suppressed tube-forming abilities of endothelial cells. PEDF and its phosphomimetic mutants achieved the antitumor effect via the inhibition of nuclear factor kappa-B (NF-κB) expression and Akt activation. The results suggest that PEDF phosphomimetic mutants may be developed as anticancer agents. 

Materials and Methods

Cells

Human choroidal melanoma cell line (OCM-1) and 293 fast-growing, highly transfectable (FT) cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Gibco-Life Technologies, Grand Island, NY) supplemented with 10% newborn calf serum and 100 U/mL penicillin/streptomycin at 37°C in a 5% CO₂ incubator. Human umbilical vessel endothelial cells (HUVEC) were cultured in DMEM with 10% fetal bovine serum and maintained at 80% confluence at 37°C in a 5% CO₂ incubator. 

Constructs and Lentiviral Vectors

Wild-type PEDF cDNA was purchased from Addgene (Cambridge, MA), whereas phosphomimetic mutants of PEDF (i.e., S24E114E227A and S24E114E227E) were constructed by site mutagenesis and confirmed by sequencing. Lentiviral expression vectors containing wild type (WT)-PEDF, S24E114E227A (EEA)-PEDF, and S24E114E227E (EEE)-PEDF were constructed by inserting the corresponding coding sequence into the multiple cloning sites (MCS) of lentiviral vector pLEX-multiple as pLEX-PEDF, pLEX-PEDF-EEA and pLEX-PEDF-EEE. pLEX-fusion green fluorescence protein-luciferase (GFP-Luc) was used as the control. To produce the lentiviruses, 293FT cells were seeded in 10-cm tissue culture dishes. When cells reached 95% confluence, the culture medium was completely removed and replaced with a fresh 10 mL of medium. Next, 800 μL of Opti-MEM containing a lentiviral packaging mix (9-μg pSPAX and 3-μg pDGM2) with 3 μg of lentiviral expression vector DNA (pLEX-PEDF, pLEX-PEDF-EEA, pLEX-PEDF-EEE or pLEX-GFP-Luc) and 800 μL of Opti-MEM containing 40 μL of lipofectamine 2000 (Invitrogen-Life Technologies, Grand Island, NY) were mixed together and applied to cells 20 minutes later. After incubation for 24 hours, the medium was replaced with a fresh medium. The lentivirus-containing medium was collected after 48 hours, centrifuged at 4000 rpm for 15 minutes, and filtered with 0.45 μm of filter before it was ready for infection or storage at −80°C. 

Stable PEDF or PEDF-Phosphomimetic Mutant Expression Cells

OCM-1 cells were infected with pLEX-PEDF, pLEX-PEDF-EEA, pLEX-PEDF-EEE or pLEX-GFP-Luc, as well as polybrene. They were then selected with medium containing puromycin (4–10 μg/mL). One week after viral infection, the remaining cells were subjected to Western blotting analysis to determine PEDF expression. 

Western Blot Analysis

For the Western blotting analysis, the OCM-1 cells, either lentivirus infected or non-infected, were washed with cold PBS twice and incubated with protein extraction buffer for 10 minutes on ice. The cell lysate was centrifuged at 13,000 rpm for 15 minutes at 4°C, and the supernatant was collected, aliquoted, and stored at −80°C. For examination of PEDF or PEDF phosphomimetic mutant expression in the cell medium, equal amounts of cells were seeded into 6-well plates. The next day, the culture medium was replaced with a medium without serum. Twenty-four hours later, the medium was harvested for immunoblotting analysis. All samples were diluted in 5× loading buffer, boiled at 95°C for 10 minutes, separated on a SDS-PAGE gel, and transferred onto polyvinylidene difluoride (PVDF) membranes. The membranes were first blocked in blocking buffer (5% milk in tris-buffered saline and Tween 20 [TBST]) at room temperature for 1 hour, then incubated with primary antibodies, such as phospho-Akt, total-Akt, PEDF, VEGF, FLK-1, PEDF receptor (PEDFR), NF-κB (p65) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Santa Cruz Biotechnology, Santa Cruz, CA), at 4°C overnight. This was followed by a 1 hour incubation with secondary antibodies, including horseradish peroxidase (HRP) conjugated goat anti-rabbit or goat anti-mouse immunoglobulin G (IgG; Santa Cruz Biotechnology). Immunoreactive bands were detected by an ECL system (Amersham Biosciences, Piscataway, NJ). 

Tube Formation Assay

An endothelial tube formation assay was performed following the manufacturer's instruction (BD Bioscience, San Jose, CA). Two hundred μL of BD Matrigel matrix (cat. No. 354,234; BD Bioscience) were added to each well of 24-well culture plates and the plates were incubated at 37°C for 1 hour to allow the gels to form. Then, 300 μL of HUVEC suspension (2 × 10⁴ cells) at low passage were added to each well. The plates were incubated for 24 hours at 37°C, 5% CO₂ atmosphere, in a medium collected from OCM-1 cells infected with either PEDF or its mutants. After 24 hours, tube formation was photographed and the number of tubes was counted. 

Cell Proliferation Assay

Cell Counting Kit-8 was used (CCK8; Wako Pure Chemical Industries, Osaka, Japan) to measure cell viability in order to evaluate cell proliferation. Both lentiviral infected and non-infected cells were seeded into 96-well culture plates in triplicate. At various time points, 10 μL of CCK-8 solution were added to each well in an assay plate and incubated in a CO₂ incubator for 1 hour. After that, 10 μL of 1% wt/vol SDS solution were added to each well to stop the reaction, and the plate was read in a microplate reader for optical density at 450 nm. 

Apoptosis Assay

Autoclaved coverslips were put into each well of 24-well culture plates before the cells were seeded. OCM-1 cells transduced with either different types of PEDF lentiviral vectors or empty vectors were immediately planted on coverslips at 70% to 80% confluency. After 72 hours, the cell media were discarded and the cells were washed with cold PBS twice, and with 500 μL of 4′,6-diamidino-2-phenylindole (DAPI) working reagent once. Five hundred μL of DAPI working reagent were added to the cells and incubated at 37°C for 15 minutes. The cells were then analyzed using a fluorescent microscope. 

Transwell Assay

Transwell chambers (Millipore-Chemicon, Billerica, MA) were used for invasion assays and migration assays. For invasion assays, matrigel coated chambers stood on empty wells of 24-well tissue culture plates. OCM-1 cells cultured in a medium without serum for 24 hours were trypsinized and suspended in a culture medium containing 0.1% BSA at 5 × 10⁴ cells/ml. Five hundred microliters of cell suspension were added into the chamber and 750 μL of complete medium were added into well. After incubation for 36 hours, the non-invading cells were removed from the upper surface of the chamber membrane with cotton tipped swabs. The cells on the lower surface of the chamber membrane were fixed in 4% PFA for 10 minutes then stained with crystal violet. The cells were counted after photographing the membrane through a microscope. For migration assays, non-coated chambers were used and migrated cells on the lower surface of the chamber membrane were fixed after a 24 hour incubation. They were then stained and photographed for further analysis. 

Wound Healing Assay

Cells were cultured in 6-well tissue culture plates and allowed to reach 95% confluence in a monolayer. A 200-μL pipette tip was used to scratch a line on the plates to create a “wound,” and the cells were allowed to grow in fresh culture medium. The photographed images of the area around the “wound” line were taken 24 and 36 hours later, and the cells in the line were counted to evaluate the healing capacity of cells cultured in media containing different types of PEDF. 

Animal Studies

All animal study protocols were approved by the local ethics committee and conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Male BALB/C nude mice were obtained from Shanghai Laboratory Animal Co. (SLAC; Shanghai, China). All mice were 6 to 8 weeks of age at the time of tumor cell injection. Three million OCM-1 cells suspended in 100 μL of BD matrix/PBS (1:1) solution were subcutaneously injected into the flank region of each mouse. Tumors were allowed to develop to a size less than 1.5 cm³. During tumor development, tumor dimensions were recorded every 2 days from the day they were visible. At the end of this experiment, the tumors were removed and weighed. The other organs, such as lungs and livers, were collected for further histological examination. 

Haematoxylin and Eosin (H&E) Staining and Immunohistochemical (IHC) Staining

The tumors were subjected to H&E and IHC stains, as previously reported. Primary antibodies used include α-smooth muscle actin (α-SMA) (Dako, Carpinteria, CA), VEGF (Santa Cruz Biotechnology), Ki67 (Santa Cruz Biotechnology), and cleaved caspase-3 (Cell Signaling Technology, Inc., Danvers, MA). 

Statistical Analysis

All quantified data were presented as a mean ± SD. The Mann-Whitney U test was used for band-intensity analysis of the Western blotting. The other data were statistically analyzed by ANOVA or Student's t-test. A P value less than 0.05 was considered statistically significant. 

Results

Effect of PEDF and its Phosphomimetic Mutants on Choroidal Melanoma Cell Proliferation and Apoptosis In Vitro

To assess the antitumor activity of WT-PEDF and its phosphomimetic mutants, we created lentiviral constructs containing cDNA derived from WT-PEDF or its phosphomemitic mutants, EEA-PEDF and EEE-PEDF, according to Grossniklaus's report and as shown in Figure 1A. OCM-1 cells were transfected with these lentiviral constructs and cultured in a selection medium supplemented with G418. After 14 days with G418 selection, 98% of OCM-1 cells were green fluorescence protein (GFP) positive (data not shown). A Western blotting analysis demonstrated that OCM-1 cells infected with lentivirus containing WT-PEDF, EEA-PEDF, or EEE-PEDF expressed high PEDF, in comparison with OCM-1 cells infected with the control lentivirus expressing a non-target gene and cells with PBS treatment (Fig. 1B). OCM-1 cells expressing PEDF and its mutants were evaluated for apoptosis by DAPI staining assay; the number of apoptotic cells in OCM-1 cells infected with one of the three PEDF lentiviral constructs was significantly higher than that of the control. In addition, OCM-1 cells infected with EEA-PEDF and EEE-PEDF showed 1.3 and 1.5 times, respectively, more apoptotic cells than those infected with WT-PEDF (Figs. 1C, 1D). We further examined whether these PEDF constructs influenced OCM-1 survival or proliferation in vitro using the CCK-8 assay (Fig. 1E). All three PEDF lentiviral constructs significantly reduced the living OCM-1 cell population (P < 0.05). However, EEA-PEDF and EEE-PEDF displayed significantly stronger effects on OCM-1 cells than WT-PEDF (P < 0.01, Fig. 1E). 

Figure 1. 

 

Comparison of PEDF and its phosphomimetic mutants on OCM-1 cell proliferation and apoptosis. (A) Schematic representation of WT-PEDF and its phosphomimetic mutants, EEA-PEDF and EEE-PEDF. (B) Western blotting probed by anti-PEDF antibody. (C) DAPI staining for morphologically changed nucleus. (D) Analysis of apoptotic cells. (E) Cell proliferation assay using CCK-8. Data is shown as mean ± SD. *P < 0.05, **P < 0.01, compared with control.

PEDF Phosphomimetic Mutants Reduce OCM-1 Invasion and Migration In Vitro

The transwell assay was employed to investigate the effect of WT-PEDF and its phosphomimetic mutants on OCM-1 cell migration and invasion. OCM-1 cells infected with either EEA-PEDF or EEE-PEDF lentiviral vectors displayed significantly reduced ability to migrate and invade in comparison with OCM-1 cells infected with empty lentiviral vectors or control cells (P < 0.05, Figs. 2A–C). OCM-1 cells infected with WT-PEDF vector also showed a trend of reduced migration and invasion ability; however, this effect was not statistically significant. Similarly, both EEA-PEDF and EEE-PEDF inhibited OCM-1 migration in the wound-healing assay (P < 0.05, Figs. 2D, 2E). 

Figure 2. 

 

Comparison of PEDF to its phosphomimetic mutants on OCM-1 cell invasion and migration in vitro. (A) The representative field of migrated cells on transwell membrane. (B, C) Analysis for cell numbers on transwell membrane from migration assay and invasion assay. (D) Differential cell migration on the scratched surface among PEDF and its phosphomimetic mutant transduced OCM-1 cells. (E) Analysis for cell migration from wound healing assay. Data is shown as mean ± SD. *P < 0.05, compared with control.

PEDF and its Phosphomimetic Mutants Impair Endothelial Cell Tube Formation In Vitro

To investigate the ability of secreted PEDFs on endothelial cell tube formation, the cell culture medium from OCM-1 cells infected with different PEDF lentiviral vectors were collected and subjected to Western blotting analysis for PEDF expression. PEDF proteins were secreted by the infected OCM-1 cells (Fig. 3A). HUVECs were cultured on a matrix (BD Biosciences) with the same amount of media as the OCM-1 cells, and tube formation was counted after 48 hours. Conditioned media containing either WT-PEDF, EEA-PEDF, or EEE-PEDF protein significantly inhibited the tube formation ability of HUVECs than media from control OCM-1 cells. Media containing EEA-PEDF and EEE-PEDF, however, displayed more profound effects than media with WT-PEDF (P < 0.01, Figs. 3B, 3C). 

Figure 3. 

 

PEDF and its phosphomimetic mutant differentially influence tube forming capacity of HUVEC. (A) Western blotting for determination of secreted PEDF in medium. (B) The representative picture for differential tube forming capacity of HUVEC cultured using medium containing PEDF or its phosphomimetic mutant. (C) Quantitative analysis for differential branch points. Data is shown as mean ± SD. *P < 0.01, compared with control.

PEDF Constructs Inhibit Tumor Formation from OCM-1 Cells In Vivo

OCM-1 cells, either mock infected or lentiviral vector infected, were implanted subcutaneous (SC) into nude mice. On the eighth day following implantation, tumors became visible in mice implanted with control OCM-1 cells. The tumors' volumes were measured every 2 days for 22 days. Mutant PEDF caused profound attenuation in tumor growth (P < 0.01, Fig. 4A). On day 22, tumor volume was highly reduced: from 1.4 cm³ (control and PBS-mock) and 0.8 cm³ (WT-PEDF) to approximately 0.3 cm³ (EEA-PEDF and EEE-PEDF) (Fig. 4B). When the mice were killed for tumor collection, each tumor's weight was measured, the results of which were consistent with the tumor volume data (Fig. 4C). To our surprise, metastasized tumor cells were detected in lungs of mice implanted with control OCM-1 cells and PBS-mock infected OCM-1 cells, but not in mice implanted with OCM-1 cells infected with the PEDF lentiviral vectors (Fig. 4D). Similarly, control OCM-1 cells, rather than lentiviral PEDF-infected OCM-1 cells, exhibited liver metastasis in the nude mice (data not shown). 

Figure 4. 

 

PEDF and its phosphomimetic mutants differentially affect OCM-1 growth and metastasis in vivo. (A) Differential tumor volume from PEDF and its phosphomimetic mutant transduced OCM-1 cells. (B) The representative picture of tumors grown in nude mice. (C) Quantitative analysis for tumor weight. Data is shown as mean ± SD. *P < 0.05, **P < 0.01, compared with control. (D) The representative picture for metastasis in lungs among untransduced and PEDF as well as its phosphomimetic mutant-transduced OCM-1 cells.

Mutant PEDF Inhibited Angiogenesis, Cell Proliferation, and Induced Apoptosis

At the end of the in vivo study, OCM-1 xenografts were sectioned and analyzed for blood vessel density using anti–α-SMA antibodies, a well recognized marker for blood vessels. Xenografts formed from OCM-1 cells expressing all three PEDF proteins exhibited lower vascular density than the xenografts from control OCM-1 cells. However, tumors expressing mutant PEDF exhibited much lower vascular density, with approximately 70% and 80% reduction in EEA-PEDF and EEE-PEDF, respectively (P < 0.01, Figs. 5A, 5B). Immunohistochemistry staining also showed that VEGF expression was much lower in tumors formed from OCM-1 cells expressing PEDF (Figs. 5A, 5C). Similarly, PEDF expression significantly increased apoptotic cell death of tumor cells, with eight times (EEA-PEDF), nine times (EEE-PEDF), and two times (WT-PEDF) more cleaved caspase-3–positive cells than tumors derived from control OCM-1 (Figs. 5A, 5D). Xenografts formed from OCM-1 cells expressing PEDF, especially EEA-PEDF and EEE-PEDF, also displayed reduced proliferation rate as demonstrated by reduced staining with Ki67, an established marker of cell proliferation (P < 0.01, Figs. 5A, 5E). 

Figure 5. 

 

Differential angiogenesis, cell proliferation and apoptosis among tumors derived from PEDF and its phosphomimetic mutant-transduced OCM-1 cells. (A) The representative picture of IHC staining from tumor section derived from PEDF and its phosphomimetic mutant-transduced OCM-1 cells. Less vasculature (α-SMA), less angiogenic factor VEGF expression, less proliferating cells (Ki67) and more apoptotic cells (cleaved caspase-3) were observed in xenografts derived from PEDF phosphomimetic mutant-transduced OCM-1 cell. (B–E) Quantitative analysis for blood vessel density, VEGF, cleaved caspase-3, and Ki67 expression. Data is shown as mean ± SD. *P < 0.05, **P < 0.01, compared with control.

To illustrate the mechanisms of the anticancer effect of PEDFs, Western blotting analyses were performed to investigate expression of proteins relevant to tumor angiogenesis. As in Figure 6A, there were no differences in protein levels of PEDFR and VEGF receptor 2 (FLK-1), but VEGF expression was slightly reduced in tumors formed from WT-PEDF cells and significantly reduced in tumors formed from EEE-PEDF and EEA-PEDF expression cells (P < 0.01, Figs. 6A, 6B). Expression of NF-κB (p65) was correspondingly reduced in tumors formed from EEE-PEDF and EEA-PEDF expression cells (P < 0.05, Figs. 6A, 6C). More interestingly, Akt activation, as evidenced by the ratio of phosphorylated Akt versus total Akt protein, was inhibited in tumors formed from OCM-1 cells expressing any of the three forms of PEDF, but more profoundly in tumors from cells expressing the mutant forms (P < 0.01, Figs. 6A, 6D). 

Figure 6. 

 

Differential gene expression in PEDF and its phosphomimetic mutant-transduced OCM-1 cells. (A) Western blot detecting gene expression in PEDF and its phosphomimetic mutant-transduced OCM-1 cells. Quantitative analysis for VEGF (B), NF-κB (C), and phospho-AKT (D) expression. Data is shown as mean ± SD. *P < 0.05, **P < 0.01, compared with control.

Discussion

PEDF is a natural antiangiogenic factor that plays a key role in the reduction of abnormal neovascularization in the eye. Its antiangiogenic activity is far greater than that of any other known endogenously produced factor. Recent studies show that PEDF is widely expressed throughout the human body and is persistent in systemic circulation. However, circulation levels of PEDF decline in patients with progressive tumors. Nevertheless, solid tumors are unable to grow over 1 mm³ without the development of new blood vessels, as their growth and expansion are thought to be totally dependent on angiogenesis. ³⁰ Thus, PEDF has been widely tested as an antiangiogenic agent to suppress tumor growth. 

Several pieces of evidence indicate that PEDF is implicated in tumor development and metastasis. First, downregulation of PEDF mRNA and/or protein has been detected in a wide range of human malignancies, ^8–11 and data suggest that PEDF expression is inversely correlated with cancer progression, ⁹ metastatic potential, ^12,13 and less favorable prognoses. ^11,12 For example, microarray studies indicated that PEDF expression is lost in highly invasive melanomas and abundant PEDF expression is restricted to the poorly aggressive counterparts. ¹⁴ Second, PEDF therapy resulted in a profound inhibition of tumor growth in animal models. ^8,16–19 However, the exact molecular mechanism by which PEDF causes tumor suppression is not fully elucidated, since the receptors of PEDF have not been completely delineated. Although a few candidates, such as extracellular matrix components, a phospholipase-linked membrane protein, ³¹ and a laminin receptor, ³² have been identified as putative receptors of PEDF, these receptors are unable to mediate all known functions of PEDF. The widely accepted hypotheses is that PEDF suppresses tumor growth via its antiangiogenic activity, and this effect is mediated through VEGF-dependent or -independent pathways that may be involved in a proapoptotic effect towards vascular endothelium cells. In addition, recent studies ^33,34 suggest that PEDF may directly affect tumor cells. For example, growth rates of human malignant melanoma G361 cells stably transfected with PEDF cDNA were significantly lower than those of the parental cells, PEDF proteins induced apoptotic cell death in cultured G361 cells in a dose-dependent manner. ³³ PEDF-induced apoptosis in A172 and U87 glioma cells was associated with increased expression of p53 and Bax, and inhibition of Bcl-2. ³³ There was a decrease in transendothelial migration in PEDF-transduced B16LS9 melanoma cells. ³⁴ PEDF also significantly reduced glioma cell migration in vitro with a significant reduction of matrix metalloproteinase-9 (MMP-9) expression. ³³ In contrast, interference with PEDF expression significantly enhanced the migratory and invasive capability of normal melanocytes and increased their proliferative potential. ³⁴ These merits make PEDF a very attractive candidate gene for tumor gene therapy. 

PEDF is a phosphoprotein: Serine 24 and 114 can be phosphorylated by casein kinase 2 (CK2), and serine on 227 by protein kinase A (PKA), which regulates its activity. ^20,21 It is postulated that more potent antitumor effects may be obtained by making phosphomimetic mutants of PEDF. Indeed, Konson et al. ²² report that triple mutants that mimic the PKA- and CK2-phosphorylated PEDF inhibit tumor growth better than WT-PEDF, and are comparable to the established antiangiogenic agent bevacizumab (Avastin; Genentech/Roche, South San Francisco, CA). They found that enhanced antitumor effects of PEDF mutants are associated with a more profound reduction in MVD and an increased ability to induce apoptosis of endothelial cells via a VEGF-independent manner. It is worth a mention that the results from this study indicate that PEDF mutants efficiently induce apoptosis of cultured endothelial cells, but do not affect the survival of cancer cells, suggesting that the antitumor effect of these agents is indirect and is achieved mainly through antiangiogenic activity. 

In order to find potent therapeutic agents for choroidal melanoma, we investigated whether PEDF triple mutants suppressed choroidal melanoma growth and metastasis in vitro and in vivo. We used lentiviral vectors to infect melanoma cells and achieve PEDF expression in tumor cells rather than using recombinant proteins. Our approach produced some different results from what has been observed by Konson et al. ²² We showed that expression of PEDF or its phosphomimetic mutants in choroidal melanoma OCM-1 cells suppressed tumor cell proliferation, migration, and invasion in vitro. OCM-1 cells expressing PEDF or its phospomimetic mutants displayed greatly reduced tumor formation ability in the nude mice. Furthermore, tumor metastasis was completely absent in nude mice implanted with OCM-1 cells expressing PEDF or its phospomimetic mutants. As expected, the phospomimetic mutants displayed stronger antitumor effects than WT-PEDF. 

To explore the underlying mechanism, we performed immunoblotting and IHC analysis of tumor cells and specimens from the xenografts. We found that tumors formed from OCM-1 cells expressing the mutant forms of PEDF have greatly reduced levels of VEGF protein. In addition, conditioned media from OCM-1 cells expressing PEDF or its phospomimetic mutants inhibited the tube formation of HUVECs. Accordingly, tumors from OCM-1 cells expressing the mutant form of PEDF also displayed much lower vascular density than tumors formed from parental OCM-1 cells. The evidence from our study indicated that PEDF or its phospomimetic mutants inhibited angiogenesis involved in VEGF-dependent pathways. In addition, the evidence from our study also indicated that PEDF or its phospomimetic mutants inhibited tumor cell proliferation, migration, and invasion and induced tumor cell apoptosis. Furthermore, OCM-1 tumor cells expressing the mutant forms of PEDF displayed inhibited Akt activation and decreased NF-κB expression. Because Akt activation and NF-κB expression are associated with stimulation of cell proliferation and inhibition of apoptosis, inhibition of Akt activation might explain the reduced proliferation signal (Ki-67) and enhanced cleaved caspase-3 activity in these tumors. ^35–37  

The differences in the experimental approaches might explain the stronger antitumor effects of PEDF and its phosphomimetic mutants from our study than those reported by Konson et al. ²² The lentiviral vector-mediated expression of PEDF in tumor cells themselves could result in continued presence of PEDF and potentially much higher local PEDF levels. Additionally, intracellular PEDF might also explain the observed anti-proliferation and pro-apoptotic effects on infected OCM-1 cells. On the other hand, choroidal melanoma cell used in the present study might be more sensitive to the antitumor effect of PEDF and its phosphomimetic mutants. However, additional studies are needed to fully elucidate these differences. 

Nevertheless, our study results are in accordance with the notion that the phosphomimetic mutants of PEDF display stronger antitumor effects than wild-type PEDF. The underlying mechanisms, both anti-angiogenesis in the tumor and anti-proliferation, and pro-apoptotic effects on tumor cells, might be mediated by reduced expression of VEGF and inhibition of Akt activation. Our results suggest that phosphomimetic mutants of PEDF provide a novel therapeutic agent for patients with melanoma. 

Footnotes

 Supported by grants from the National Basic Research Program of China (2010CB529902), the Science and Technology Committee of Shanghai (064119539), and Leading Medical Talent of Shanghai (040308).

Footnotes

 Disclosure: Y. Feng, None; W. Bao, None; Y. Luo, None; L. Tian, None; X. Chen, None; M. Yi, None; H. Xiong, None; Q. Huang, None